Active appearance models with occlusion
Active appearance models with occlusion
复制标题
DOI：
10.1016/j.imavis.2005.08.001
复制
发表时间：
2006-06-01
影响因子：
4.7
通讯作者：
Baker, Simon
中科院分区：
文献类型：
--
作者：
Gross, Ralph;Matthews, Iain;Baker, Simon
主动外观模型(AAM)是一种生成性参数模型，过去已成功地用于跟踪视频中的人脸。各种视频应用是可能的，包括用于实时用户界面的动态头部姿势和凝视估计。唇读和表情识别。为了构建AAM，需要一些带有典型特征点(通常是手工标记的)网格的人脸的训练图像。所有特征点必须在所有训练图像中可见。然而，在许多情况下，面部的某些部分可能会被遮挡。也许最常见的遮挡原因是3D姿势变化，这可能会导致面部的自我遮挡。更重要的是。在存在微小遮挡的情况下，使用标准AAM拟合算法的跟踪经常失败。在本文中，我们提出了从遮挡的训练图像中构造AAM的算法，并在包含遮挡的视频中有效地跟踪人脸。我们对我们的算法进行了定量和定性的评估，并在包含不同程度和类型的遮挡的多个图像序列上成功地进行了实时人脸跟踪。(C)2005 Elsevier B.V.保留所有权利。
Active Appearance Models (AAMs) are generative parametric models that have been successfully used in the past to track faces in video. A variety of video applications are possible, including dynamic head pose and gaze estimation for real-time user interfaces. lip-reading, and expression recognition. To construct an AAM, a number of training images of faces with a mesh of canonical feature points (usually hand-marked) are needed. All feature points have to be visible in all training images. However, in many scenarios parts of the face may be occluded. Perhaps the most common cause of occlusion is 3D pose variation, which can cause self-occlusion of the face. Furthermore. tracking using standard AAM fitting algorithms often fails in the presence of even small occlusions. In this paper we propose algorithms to construct AAMs form occluded training images and to track faces efficiently in videos containing occlusion. We evaluate our algorithms both quantitatively and qualitatively and show successful real-time face tracking on a number of image sequences containing varying degrees and types of occlusions. (c) 2005 Elsevier B.V. All rights reserved.